CTRL-Rec: Controlling Recommender Systems With Natural Language

2025-10-14

AI Summary

Existing recommender systems lack fine-grained, real-time, user-controllable mechanisms. To address this, we propose CTRL-Rec, a method that distills the natural language understanding capability of large language models (LLMs) into a lightweight embedding model, enabling users to dynamically express preferences via free-text input and instantaneously refine recommendations. CTRL-Rec requires only a single LLM embedding computation per query, ensuring low-latency response and seamless integration into conventional recommendation architectures. Extensive evaluation on the MovieLens dataset demonstrates its effectiveness, while an in-the-wild study with 19 Letterboxd users shows a 37.2% increase in perceived user control and significantly higher overall satisfaction compared to baselines (p < 0.01). Our core contribution is the first high-fidelity distillation of LLM semantic capabilities into an efficient, production-ready recommendation control module—achieving both expressive flexibility and practical deployability.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

When users are dissatisfied with recommendations from a recommender system, they often lack fine-grained controls for changing them. Large language models (LLMs) offer a solution by allowing users to guide their recommendations through natural language requests (e.g.,"I want to see respectful posts with a different perspective than mine"). We propose a method, CTRL-Rec, that allows for natural language control of traditional recommender systems in real-time with computational efficiency. Specifically, at training time, we use an LLM to simulate whether users would approve of items based on their language requests, and we train embedding models that approximate such simulated judgments. We then integrate these user-request-based predictions into the standard weighting of signals that traditional recommender systems optimize. At deployment time, we require only a single LLM embedding computation per user request, allowing for real-time control of recommendations. In experiments with the MovieLens dataset, our method consistently allows for fine-grained control across a diversity of requests. In a study with 19 Letterboxd users, we find that CTRL-Rec was positively received by users and significantly enhanced users'sense of control and satisfaction with recommendations compared to traditional controls.

Problem

• Enabling natural language control for recommender systems
• Addressing user dissatisfaction with fine-grained recommendation adjustments
• Integrating LLM-based predictions into traditional recommendation algorithms

Innovation

• Natural language control for recommender systems
• LLM-simulated user approval for training embeddings
• Real-time integration of user requests into recommendations